The diversity, classification and historical biogeography of the mayfly family Prosopistomatidae are explored. First, the higher classification of the Ephemeroptera is reviewed, focussing on the phylogenetic placement of the Prosopistomatidae relative to other mayfly families. All relevant literature from 1762 to 2010 is synthesized. Baetiscidae are established as the probable sister lineage of Prosopistomatidae, the two constituting the superfamily Baetiscoidea. Next, qualitative morphological variation within the Prosopistomatidae is reviewed and revised, emphasizing nymphs because imaginal specimens are few. The labium and associated structures and the hypopharynx of nymphs, and the highly-derived wing venation of the imaginal stages, are re-interpreted. The structure of the male tarsal claws changes considerably between subimago and imago, which, together with deeply scalloped ridges on male imaginal forelegs and unusual pits on the female thorax, are interpreted as providing an unusual mating mechanism. These structures provide morphological characters for species definition and phylogenetic analyses. Two approaches to species delimitation are explored. First, morphometric variation is analysed using Principal Component Analysis, revealing groupings that can be interpreted as species, although there is some overlap between them. Discriminant Function Analysis shows that head width and carapace shape have the most value in identifying nymphs of different species. The carapace of Prosopistoma nymphs is shown to grow allometrically and gradually, in contrast with that of Baetisca, indicating a difference in early ontogeny. Second, an Artificial Neural Network algorithm applied to nymphal morphological characters accurately identified species. This computer-driven artificial intelligence method has power to provide future easy-to-use electronic identification aids. Phylogenetic analysis of nymphal morphology using the parsimony method shows two clades of Prosopistomatidae, one sharing characters with the type species, Prosopistoma variegatum and the other predominating in Africa, although also occurring in Asia; these clades are named the “P. variegatum” and “African” clades, respectively. Parsimony analysis of adult morphology supports these two clades, but supertree analysis obscures the relationships, nesting the “P. variegatum” lineage within the other clade. Preliminary molecular phylogenetic analysis of the 16S rRNA, (mitochondrial) 18S rRNA and Histone-3 genes using Bayesian Inference methods does not support the two clades shown by morphology. Instead, there is a strong relationship between the European species and one African species, with the single Asian representative being most distantly related. These results are limited by lack of fresh material, patchy taxon sampling, and problems with finding suitable primers. A molecular clock program, BEAST, calibrated using fossils, suggests divergence times for the oldest crown-group Prosopistoma clade, represented by the Asian P. wouterae, of about 131 Ma, with the youngest species, the African P. crassi, of 1.21 Ma. Stem-group relationships are analysed using parsimony analysis, focussing on wing characters of the Baetiscoidea, other extant mayfly lineages, and extinct stem-group lineages. This suggests that the Baetiscoidea diverged from main-line Ephemeroptera earlier than any other extant mayfly lineage. This approach expands upon ideas hinted at by earlier scientists. Finally, historical biogeographical analysis of the distribution of known Baetiscoidea s.s. stem-group fossils implies a once Pangean distribution of the lineage. Changing palaeo-climate, catastrophic extinction events and plate tectonic movements in relation to the distribution of crown-group species are reviewed. Other approaches to historical biogeography that build on both morphological and molecular phylogenies are used to interpret disperalist and vacarianist arguments. Distribution patterns of eight unrelated freshwater organisms which share a similar distribution pattern are compared, assuming that shared patterns indicate similar historic biogeographic processes. The distribution of recent Prosopistoma species is seen to be the product of evolution resulting from both vicariance and dispersal. In conclusion, this thesis encompasses a variety of disciplines. It successfully recognises new characters and distinguishes previously unknown species. It uses new approaches to delimiting species and known methods to determine phylogeny from several angles. The analysis of stem-group relationships offers an insight into possible early lineage splitting within Ephemeroptera. Interpretation of historical biogeography allows for both a Gondwanan origin of Prosopistomatidae, with rafting of species on the Deccan plate to Asia, and for subsequent dispersal from Asia down to Australia and across to Europe, and possibly back to Africa.